- Developed from scratch a service that improves the accuracy of seismic trace processing by up to 93%. - Implemented a GUI that made the generation of training data for ML models 8 times faster. - Optimized and expanded a service that improves the resolution of processed seismic images, reducing processing cycle times by more than 5 hours per week. - Presented work at 2 international scientifi c conferences and workshops with 3 major oil and gas companies. Skills: - Python (Jupyter, NumPy, TensorFlow, PyTorch, SciPy, plotly); - Linux (Bash); - git.

- Launched 3 trading strategies that improved trading performance and profi tability for the clients. - Proposed custom solutions focused on automating parameter changes based on predefi ned thresholds rather than manual processes, resulting in a savings of 20 hours per month. - Engineered robust scheduled data processing pipelines, achieving a 95% reduction in human error rates. Skills: - Python 3.10 (jupyter, numpy, pandas, matplotlib, scipy, sklearn, catboost, dagster); - Linux (Bash); - git; - AWS; - cron.

- Designed a mathematical model that increased client revenue by approximately 40%. - Orchestrated the creation of a robust microservices architecture connecting complex mathematical algorithms to external service providers. This integration streamlined processes by consolidating 8 disparate systems into one cohesive solution. - Optimized computation time that increased the bandwidth of the service by approximately 50 times. Технологический стек: - Python 3.9 (jupyter, numpy, pandas, matplotlib, scipy, xlwings, sklearn, pymongo, locust); - git; - MongoDB; - MS SQL; - gRPC; - Docker; - Microsoft Azure.

- Authored a comprehensive Python framework specifi cally designed to simulate heating and cooling phenomena using molecular dynamics. Use of this framework has resulted in 7 scientifi c publications. - Implemented advanced machine learning models that predict structural properties of systems at diff erent temperatures, reducing the computational time by more than 15 times compared to classical methods. - Presented fi ndings and insights at 6 international conferences, secured publication of research in a peer-reviewed journal. Skills: - Python 3.10 (numpy, scipy, matplotlib, numba, sklearn, keras, jupyter); - git; - LaTeX.

- Upgraded 2 probabilistic process models that improved predictive accuracy and decision-making capabilities. - Developed unique tools that resulted in the collection of approximately 10,000 records per month, ensuring continuous availability of relevant input parameters essential for advancing probabilistic analyses with minimal downtime. - Authored detailed model implementation documentation that enhanced knowledge sharing among team members, leading to a 60% reduction in onboarding time for new projects. Skills: - Python 3.8 (jupyter, numpy, pandas, matplotlib, scipy, xlwings, beautifulsoup4, selenium); - git.

- Created a Python package to calculate the crystalline electric fi eld parameters of rare earth ions in solids, accelerating research by 10 hours per week. - Streamlined the initial parameter fi tting process by automating procedures, reducing setup time from 60 to 5 minutes per experiment without compromising data quality. - Leveraged advanced mathematical techniques using Python's NumPy and SciPy frameworks to achieve computational accuracy for solid-state physics research, resulting in a 70% improvement in data reliability. Skills: - Python 3.7 (numpy, scipy, matplotlib); - git.